The friction log: capturing what your week is telling you

A friction log is a running list of every moment your work dragged: the file you couldn't find, the decision you made for the third time, the client email that took 40 minutes because there was no template. One line per entry, written at the moment of annoyance, reviewed weekly. Kept for a month, it becomes the most honest improvement backlog you own — better than any brainstorm, because it is built from evidence rather than opinion.

The log is the sensing layer of my personal operating system: the part that notices problems so the rest of the system can fix them.

Why write friction down at all?

Because friction is information, and untracked information evaporates. Every founder experiences dozens of small drags per week; almost none are memorable individually, so the week ends with a vague sense of busyness and no record of where it went. The same drag then repeats next week, and the week after — small enough to tolerate each time, expensive in aggregate.

Writing it down converts irritation into data. A single entry means nothing. The same entry appearing four times in a month is a system defect with a frequency, and frequency is what tells you it is worth fixing.

What counts as friction?

Anything that made work slower, heavier or more annoying than it should have been. In practice, entries cluster into a few types:

• Repeated decisions — choosing the same thing again ("what do I charge for X?", "which slot do I offer?").
• Repeated writing — composing near-identical emails, proposals or answers from scratch.
• Searching — hunting for files, links, logins, or "that thing the client sent".
• Waiting and chasing — progress blocked on someone else, with no reminder system doing the chasing.
• Energy drains — tasks or meetings that leave you flattened; these overlap with managing energy as the real constraint and deserve logging even when they are technically "fine".

The bar for an entry is deliberately low. If you noticed it, log it. Filtering happens at review, not at capture — a capture step with a quality gate stops getting used within a week.

How does the log actually work?

The mechanism is short:

1. When something drags, then write one line in a single fixed place — one note file, one pocket notebook, one voice memo thread. The place matters more than the tool; two capture points means zero reliable capture points.
2. When you write the line, then include what happened and roughly what it cost ("rewrote scope email from scratch, ~30 min"). No analysis at capture time.
3. When the week ends, then read the log once and tally repeats. This takes five minutes inside a weekly review.
4. When an item has appeared three or more times, then it graduates to the fix list: template it, automate it, delegate it, or decide it once and write the rule down.
5. When a fix ships, then the entries stop appearing — which is the log confirming the fix worked, at no extra measurement cost.

The whole system costs perhaps ten minutes a week. Its output is a prioritised, evidence-based list of what to systematise next.

How do you pick which friction to fix?

By frequency times cost, roughly. A 30-minute drag that appears weekly costs you around 25 hours a year; a mild daily one can cost more. Fix the ones with the biggest product first, and prefer fixes that delete the task over fixes that speed it up.

One caution from the building side: not everything that repeats deserves automation. Some fixes are a template; some are a sentence in your proposal; some are a price. When clients ask me to automate a process, the friction log question comes first — how often, at what cost — because automation carries maintenance debt of its own. It is the same logic that makes me argue hourly billing for automation work is a trap: if nobody has measured the recurring cost, nobody can judge what fixing it is worth.

What changes after a month?

Two things. First, you ship a handful of small fixes that permanently remove drag — templates, rules, checklists, the odd automation. Second, and more useful, you stop trusting your intuition about where the week goes, because the log keeps contradicting it. The tasks founders complain about are rarely the ones bleeding the most time; the log finds the quiet ones.

Start today: one file, one line per annoyance, first review on Friday. The week is already telling you what to fix. The log just writes it down.

---

Keep the thinking coming: The Format — a fortnightly letter on business and personal systems. Subscribe by email.